ZeroSlop — June 11, 2026
Today: HERO: Hindsight-Enhanced Reflection from Environment…; Google DeepMind is worried about what happens when…; “Giant superatoms” could finally solve quantum…
12 stories worth knowing about today — AI breakthroughs, launches, and innovations making a difference.
1. HERO: Hindsight-Enhanced Reflection from Environment Observations for Agentic Self-Distillation
arXiv CS.AI
arXiv:2606.11559v1 Announce Type: new Abstract: Reinforcement learning typically improves multi-turn agent capabilities through the terminal outcome of the trajectories, which makes it difficult to determine credit assignments for each intermediate turns. Recent on-policy self-distillation methods …
2. Google DeepMind is worried about what happens when millions of agents start to interact
MIT Tech Review
Google DeepMind is funding research into the potential dangers of situations where millions of different AI agents interact with each other online. According to Rohin Shah, who directs the company’s AGI safety and alignment research, the mass-market arrival of agents that can carry out tasks without…
3. “Giant superatoms” could finally solve quantum computing’s biggest problem
Science Daily AI
In the pursuit of powerful and stable quantum computers, researchers at Chalmers University of Technology, Sweden, have developed the theory for an entirely new quantum system – based on the novel concept of ‘giant superatoms’. This breakthrough enables quantum information to be protected, controlle…
4. AGI Is Not Multimodal
The Gradient
“In projecting language back as the model for thought, we lose sight of the tacit embodied understanding that undergirds our intelligence.” –Terry Winograd The recent successes of generative AI models have convinced some that AGI is imminent. While these models appear to capture the essence of human…
5. Running Python code in a sandbox with MicroPython and WASM
Simon Willison
I’ve been experimenting with different approaches to running code in a sandbox for several years now, but my latest attempt feels like it might finally have all of the characteristics I’ve been looking for. I’ve released it as an alpha package called micropython-wasm , and I’m using it for a code ex…
6. Jedify raises $24M to help companies arm AI agents with context on their business
TechCrunch AI
The funding round was led by Norwest, with participation from S Capital VC, Cerca Partners, and Oceans Ventures. Snowflake Ventures also participated as a strategic investor….
7. For Robotaxis, Safety Must Be Built In, Not Bolted On
NVIDIA Blog
A car pulls up to the curb. The app says, “Your ride is here.” No one’s in the driver’s seat. For people who live in one of the dozens of cities now hosting robotaxi services, this is already a reality. The robotaxi industry has moved from prototype milestones to commercial operations, with an expan…
8. Architecture-Aware Reinforcement Learning Makes Sliding-Window Attention Competitive in Math Reasoning
arXiv CS.AI
arXiv:2606.11634v1 Announce Type: new Abstract: The rapid progress of reasoning and agentic large language models (LLMs) has increased the demand for long-context inference, but self-attention (SA) scales quadratically with context length. To address this, we study SWARR (Sliding-Window Attention…
9. ChatGPT as a therapist? New study reveals serious ethical risks
Science Daily AI
As millions turn to ChatGPT and other AI chatbots for therapy-style advice, new research from Brown University raises a serious red flag: even when instructed to act like trained therapists, these systems routinely break core ethical standards of mental health care. In side-by-side evaluations with …
10. DiffusionGemma
Simon Willison
DiffusionGemma Last May Google briefly released an experimental Gemini Diffusion model. I tried the preview at the time and recorded it running at 857 tokens/second. It was an exciting model, but Google made no further announcements about it. That research has returned in the best possible way: as a…
11. RVPO: Risk-Sensitive Alignment via Variance Regularization
Apple ML Journal
Current critic-less RLHF methods aggregate multi-objective rewards via an arithmetic mean, leaving them vulnerable to constraint neglect: high-magnitude success in one objective can numerically offset critical failures in others (e.g., safety or formatting), masking low-performing “bottleneck” rewar…
12. Claude Fable won’t answer basic biology questions
The Verge AI
Anthropic just released Claude Fable 5, calling it the most powerful AI model it has ever made widely available and praising its skills in biology, among others. But the model won’t answer basic biology questions - the kind you’d expect a high schooler to handle. Instead, it hands off the query to t…